gazelle mobile eye-tracking for sportsmobile eye-tracking for sports reto meier, damian weber, ralf...

1
source: https://doi.org/10.7892/boris.68986 | downloaded: 21.3.2021 Contact: [email protected] [email protected] Quellgasse 21 | CH-2502 Biel huce.ti.bfh.ch Institute for Human Centered Engineering HuCE Gazelle Mobile Eye-tracking for Sports Reto Meier, Damian Weber, Ralf Kredel, Ernst-Joachim Hossner*, Marcel Jacomet, Theo Kluter Institute for Human Centered Engineering HuCE Institute of Sport Science ISPW, University of Bern ABSTRACT In collaboration with the Institute of Sport Science (IPSW) at the University of Bern, the microLab research group at the HuCE institute develops an outdoor eye-tracking system with scene overlay. Due to its field of application, system miniaturization and processing speed are the key elements of this project. SYSTEM & FACTS The main features of the Gazelle system include: High mobility for the use in practical sport settings Calibration-free setup due to 3D eye modeling. High frame rate for scene and eye video EYE MOVEMENTS IN SPORTS Currently, mobile eye-tracking systems are used in quasi-static environments, for example in studies for product marketing in the supermarkets. Moreover cur- rent eye tracking systems restrict the field of view and have a poor wearing comfort. Those facts make eye-trackers hardly usable in outdoor sport environments. OUR CONCEPT Eye tracking systems analyze the pupil movements and com- bine them with a scene image to visualize the wearer's gaze path. The challenge in optical systems is finding the best possible camera angle with respect to the eye, without limit- ing the field of view. Our concept is based on safety glasses for sports, in which we integrate miniature cameras. These cameras allow a direct view on the eyes from inside of the glasses. In addition, the ambient light at visible spectrum is canceled out resulting in a increased gaze tracking stability and thus better system robustness under challenging light conditions. REQUIREMENTS FOR A NEW SYSTEM To be able to compare the results obtained in lab conditions to real sports situations, the cooperation project aims to develop a highly mobile eye-tracker suitable for outdoor com- petitive sports. Light weight as well as compactness in addition to lower risk of injury are the core requirements to the new eye-tracking system. In addition, the athlete wear- ing the device should in no way be influenced by the eye-tracking system. Figure 1: Starburst algo- rithm for pupil tracking Figure 4: The new Gazelle system concept In particularly dynamic situations it can come to shifts of the eye tracking glasses. Usually this requires a re-calibration of the system. Our solution with two cameras per eye and the corresponding signal processing algorithms can create a 3D model of the eyeball and thus detect displacements of the system and compensate them in real time. In addition, the modeling allows an autonomous calibration, which means a significant increase in operator convenience compared to ex- isting eye-tracking systems. Figure 3: Our current Gazelle glasses prototype. The miniature cam- eras are fully integrated in the frame. The team at the ISPW Bern has several years of experience in researching the impact of gaze behavior on decision- making in sports situations. The eye-tracking systems used therein are mobile, but are primarily designed for lab use. Figure 2: A mountain biker wear- ing a state-of-the-art third party eye-tracker which is not appropri- ate for sports *Ordinarius für Bewegungs- und Trainingswissenschaft am Institut für Sportwissenschaft der Univeristät Bern/Schweiz

Upload: others

Post on 17-Oct-2020

4 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Gazelle Mobile Eye-tracking for SportsMobile Eye-tracking for Sports Reto Meier, Damian Weber, Ralf Kredel, Ernst-Joachim Hossner*, Marcel Jacomet, Theo Kluter Institute for Human

source: https://doi.org/10.7892/boris.68986 | downloaded: 21.3.2021

Contact: [email protected]@bfh.chQuellgasse 21 | CH-2502 Bielhuce.ti.bfh.ch

▶ Institute for Human Centered Engineering HuCE

Gazelle Mobile Eye-tracking for Sports

Reto Meier, Damian Weber, Ralf Kredel, Ernst-Joachim Hossner*, Marcel Jacomet, Theo Kluter

Institute for Human Centered Engineering HuCE Institute of Sport Science ISPW, University of Bern

ABSTRACT

In collaboration with the Institute of Sport Science (IPSW) at the University of Bern, the microLab research group at the HuCE institute develops an outdoor eye-tracking system with scene overlay. Due to its field of application, system miniaturization and processing speed are the key elements of this project.

SYSTEM & FACTS

The main features of the Gazelle system include:▶ High mobility for the use in practical sport settings▶ Calibration-free setup due to 3D eye modeling.▶ High frame rate for scene and eye video

EYE MOVEMENTS IN SPORTS

Currently, mobile eye-tracking systems are used in quasi-static environments, for example in studies for product marketing in the supermarkets. Moreover cur-rent eye tracking systems restrict the field of view and have a poor wearing comfort. Those facts make eye-trackers hardly usable in outdoor sport environments.

OUR CONCEPT

Eye tracking systems analyze the pupil movements and com-bine them with a scene image to visualize the wearer's gaze path. The challenge in optical systems is finding the best possible camera angle with respect to the eye, without limit-ing the field of view. Our concept is based on safety glasses for sports, in which we integrate miniature cameras. These cameras allow a direct view on the eyes from inside of the glasses. In addition, the ambient light at visible spectrum is canceled out resulting in a increased gaze tracking stability and thus better system robustness under challenging light conditions.

REQUIREMENTS FOR A NEW SYSTEM

To be able to compare the results obtained in lab conditions to real sports situations, the cooperation project aims to develop a highly mobile eye-tracker suitable for outdoor com-petitive sports.

Light weight as well as compactness in addition to lower risk of injury are the core requirements to the new eye-tracking system. In addition, the athlete wear-ing the device should in no way be influenced by the eye-tracking system.

Figure 1: Starburst algo-rithm for pupil tracking

Figure 4: The new Gazelle system concept

In particularly dynamic situations it can come to shifts of the eye tracking glasses. Usually this requires a re-calibration of the system. Our solution with two cameras per eye and the corresponding signal processing algorithms can create a 3D model of the eyeball and thus detect displacements of the system and compensate them in real time. In addition, the modeling allows an autonomous calibration, which means a significant increase in operator convenience compared to ex-isting eye-tracking systems.

Figure 3: Our current Gazelle glasses prototype. The miniature cam-eras are fully integrated in the frame.

The team at the ISPW Bern has several years of experience in researching the impact of gaze behavior on decision-making in sports situations. The eye-tracking systems used therein are mobile, but are primarily designed for lab use.

Figure 2: A mountain biker wear-ing a state-of-the-art third party eye-tracker which is not appropri-ate for sports

*Ordinarius für Bewegungs- und Trainingswissenschaft am Institut für Sportwissenschaft der Univeristät Bern/Schweiz